Mechanisms of spindle positioning: focus on flies and worms

332

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

Mechanisms of spindle positioning: focus on flies and worms Pierre Gönczy Accurate spindle positioning is crucial for spatial control of cell division. During metazoan development, coordination between polarity cues and spindle position also ensures correct segregation of cell fate determinants. Converging evidence indicates that spindle positioning is achieved through interactions between cortical anchors and the plus ends of microtubules, generating pulling forces acting on spindle poles. This article discusses recent findings that indicate how this mechanism might be used for spindle positioning during Drosophila and Caenorhabditis elegans development.

Pierre Gönczy Swiss Institute for Experimental Cancer Research (ISREC), 155 ch. des Boveresses, CH-1066 Epalinges/ Lausanne, Switzerland. e-mail: Pierre.Gonczy@ isrec.unil.ch

The position of the mitotic spindle plays a key role in spatial control of cell division. In animal cells, the position of the spindle at anaphase determines the placement of the cleavage furrow, which forms so as to bisect the mitotic spindle [1]. Therefore, spindle position dictates the relative size of daughter cells, because a centrally located spindle leads to equal cleavage, whereas an eccentrically located spindle results in unequal cleavage (Fig. 1a). Such asymmetric divisions are prevalent during metazoan development, when they contribute to the generation of cell fate diversity. Moreover, spindle position dictates the spatial relationship between daughter cells. For example, a spindle oriented parallel to the plane of an epithelial sheath leads to both daughter cells remaining in the epithelium, whereas a spindle oriented perpendicular to that plane results in one daughter departing from the epithelium (Fig. 1b). Such perpendicular divisions take place when, for instance, neuroblasts delaminate from the neurectodermal sheath during insect neurogenesis. Importantly, spindle position must be coordinated with cell polarity. Several components are unevenly distributed in polarized cells and spindle position dictates their inheritance by each daughter cell (Fig. 1b). Thus, polarity cues and spindle position together ensure the correct segregation of cell-fate determinants during development. What are the mechanisms that ensure proper spindle positioning? Evidence from several systems indicates that interactions between microtubules emanating from the spindle poles and the cell cortex play a major role. Here, I focus on spindle positioning in the fruit fly Drosophila melanogaster and the nematode Caenorhabditis elegans. I begin, however, by considering a general mechanistic framework for spindle positioning that draws on findings from several organisms. I end by speculating that a spindle positioning checkpoint in metazoan organisms might ensure coordination between temporal and spatial cell division control. http://tcb.trends.com

Mechanistic framework for spindle positioning

The position of the spindle follows from that of the two spindle poles, which nucleate astral and spindle microtubules of the mitotic apparatus. What are the mechanical forces that act on spindle poles to achieve proper spindle positioning? In principle, ‘pushing’ forces resulting from polymerization of astral microtubules could help to position spindle poles. When the plus end of a growing microtubule encounters a fixed object like the cortex, the addition of tubulin dimers displaces the microtubule in the opposite direction. As a result, the tethered spindle pole is also displaced away from the cortex. Although microtubule polymerization forces can position an aster in vitro or the interphase spindle-pole body in Saccharomyces cerevisiae [2,3], it remains to be established whether they contribute to the positioning of spindle poles during mitosis. By contrast, the contribution of ‘pulling’ forces that bring spindle poles closer to the cell cortex has been well documented. For instance, when the meiotic spindle is displaced from the cortex of Chaetopterus oocytes using a microneedle and then released, one of the spindle poles is rapidly pulled back towards the cortex [4]. Pulling forces also act on spindle poles in somatic cells. For example, when the spindle of rat kangaroo Ptk2 cells is severed using a laser microbeam, spindle poles move towards the cell cortex with higher peak velocities than in untreated cells, demonstrating the existence of forces external to the spindle pulling on spindle poles [5]. Findings in several systems indicate that such pulling forces are microtubule dependent. For instance, astral microtubules connect the moving spindle pole and the cortex in Chaetopterus oocytes, whereas severing of astral microtubule connections between the spindle pole and the cortex abolishes spindle pole movement in Ptk2 cells [4,5]. How are microtubule-dependent pulling forces generated? Two possibilities are suggested by an analogy between spindle positioning and chromosome segregation (Fig. 2) [6,7]. In both cases, the minus ends of microtubules are at the spindle pole, whereas the plus-ends are at either the cell cortex or the kinetochore. Therefore, microtubuledependent pulling forces driving spindle positioning might be mechanistically related to those driving chromosome segregation. In the case of spindle positioning, these forces pull the spindle pole towards the cortex because the cortex is in a fixed

0962-8924/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0962-8924(02)02306-1

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

333

(a) Cell cortex

+

–

+

(b)

TRENDS in Cell Biology

Fig. 1. The position of the spindle at anaphase dictates the relative sizes of and the spatial relationship between daughter cells. Spindle poles are shown as black disks, astral and spindle microtubules as black lines, chromosomes and nuclei as grey ovals and disks, the cleavage furrow as a dashed red line, and cell-fate determinants as blue disks. (a) A centrally located spindle results in two daughter cells of equal sizes (left), whereas an eccentrically positioned spindle results in two daughter of different sizes (right). (b) A spindle oriented parallel to the epithelial sheath results in both daughter cells remaining in the epithelium (left), whereas a spindle oriented perpendicular to the epithelial sheath results in one of the daughters leaving the epithelium (right). Spindle orientation dictates the inheritance of cell fate determinants by daughter cells. If cell fate determinants are present on the basal side of the mother cell then, when spindle orientation is parallel to the epithelium (left), both daughters inherit cell fate determinants. However, only one daughter cell inherits them when spindle orientation is perpendicular to the epithelium (right).

position. By contrast, in the case of chromosome segregation, these forces pull the kinetochore towards the spindle pole because the spindle pole and associated astral microtubules have a larger drag than the kinetochore. Two types of mechanisms are thought to generate pulling forces during chromosome segregation: microtubule-depolymerization-coupled movements and minus-end-directed motor-driven movements [8]. Support for the first of these comes, for instance, from observations that kinesin coupled to beads remains attached to the plus ends of depolymerizing microtubules in vitro [9]. Support for the second comes from the presence of the minus-enddirected motor protein cytoplasmic dynein at kinetochores in several species, as well as from the requirement for dynein for proper chromosome segregation in Drosophila [10–13]. These two mechanisms are likely to be coupled, because cytoplasmic dynein could also maintain attachment with the plus ends of depolymerizing microtubules, whereas microtubules must shorten concomitant http://tcb.trends.com

TRENDS in Cell Biology

Fig. 2. Analogous mechanisms might be used to generate pulling forces during chromosome segregation and spindle positioning. The polarities of astral and spindle microtubules are indicated, and depolymerizing microtubules are indicated by the loss of tubulin dimers from the plus ends. For simplicity, only a limited cortical area and one set of sister chromatids is represented. The spindle pole is shown as a black disk, astral and spindle microtubules as black lines, and chromosomes as a gray oval. Orange squares indicate anchors that recognize the plus ends of microtubules or minus-end-directed motors such as cytoplasmic dynein. Orange arrows indicate direction of movement resulting from pulling forces acting along microtubules. Two types of mechanisms might generate pulling forces: microtubuledepolymerization-coupled movements and motor-protein-driven movements. In the first scenario, anchors at kinetochores or the cell cortex (orange squares) maintain interactions with the plus ends of depolymerizing microtubules, thus generating pulling forces (orange arrows). In the second scenario, minus-end-directed motors such as cytoplasmic dynein anchored at kinetochores or at the cell cortex (orange squares) exhibit minus-end-directed motility, thus generating pulling forces (orange arrows). Microtubules also shorten in the second scenario, although without resulting in force generation.

with dynein moving towards the spindle pole by virtue of its motor activity. Also, dynein might act by modulating microtubule dynamics, a possibility suggested by the observation that microtubules are more stable in dynein heavy-chain mutants in S. cerevisiae and Aspergillus nidulans [14,15]. By analogy to chromosome segregation, the two types of mechanism could also generate pulling forces during spindle positioning. Cortical anchors could maintain interactions with the plus ends of depolymerizing microtubules, whereas cortically bound cytoplasmic dynein could exhibit

334

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

Box 1. How else could minus-end-directed motors exert pulling forces on spindle poles?

(a)

C. elegans

(b)

Drosophila

The prevailing view for the generation of pulling forces acting on spindle poles invokes interactions between microtubules and the cell cortex. However, this need not be the case. For instance, minus-end-directed motors such as cytoplasmic dynein could be anchored to a peripheral structure some distance from the cortex. Alternatively, minus-end-directed motors anchored throughout the cell could generate pulling forces proportional to the length of microtubules. Such lengthdependent forces have been revealed in newly fertilized sand dollar eggs [a] and could account for centrosome separation in vertebrate somatic cells [b]. In such scenarios, minus-end-directed motors must be anchored to a structure that has a higher drag than that of spindle poles; this structure might, for instance, be the actin cytoskeleton or the endoplasmic reticulum. References a Hamaguchi, M. S. and Hiramoto, Y. (1986) Analysis of the role of astral rays in pronuclear migration by the colcemid–UV method. Dev. Growth Differ. 28, 143–156 b Waters, J. C. et al. (1993) The force-producing mechanism for centrosome separation during spindle formation in vertebrates is intrinsic to each aster. J. Cell Biol. 122, 361–372

minus-end-directed motility. In both types of mechanism, pulling forces would move the minus ends of microtubules and tethered spindle poles towards the cell cortex. Although alternative mechanisms can be envisaged for the involvement of minus-end-directed motors in generating pulling forces on spindle poles (Box 1), analysis of spindle positioning in S. cerevisiae confirms that interactions between astral microtubules and cortical anchors play a major role. The mechanisms of spindle positioning in budding yeast have recently been reviewed elsewhere [16,17] and only principal findings that help in the discussion of spindle positioning in Drosophila and C. elegans are mentioned here. In S. cerevisiae, the intranuclear spindle must be correctly positioned at the bud neck to segregate chromosomes faithfully to daughter cells. Spindle positioning is achieved in two steps. First, the nucleus is positioned along the mother–bud axis in a step that requires the function of Kar9p. Kar9p is located primarily at the bud cortex and interacts with Bim1p, a protein present at the plus ends of microtubules [18–21]. Cells with a kar9∆ or a bim1∆ mutation exhibit decreased pulling of the spindle pole body towards the cell cortex and inefficient nuclear positioning [19–21]. In a second step, the nucleus moves through the bud neck in a dynein-dependent manner. The cortical protein Num1p also plays a key role in this step: components of the dynein complex and Num1p immunoprecipitate together, and num1 mutants resemble cells lacking dynein function [22–24]. Taken together, these considerations and findings support a mechanistic framework in which spindle positioning is achieved through interactions between cortical anchors and the plus ends of microtubules that result in the generation of pulling forces acting on spindle poles. Below, I discuss spindle positioning during Drosophila and C. elegans development in the light of this mechanistic framework. http://tcb.trends.com

TRENDS in Cell Biology

Fig. 3. Live imaging of spindle positioning in one-cell-stage Caenorhabditis elegans embryos and Drosophila sensory organ precursors. Elapsed time is indicated for each image in minutes and seconds. (a) Time-lapse fluorescence microscopy of one-cell-stage C. elegans embryo carrying a fusion between green fluorescent protein (GFP) and β-tubulin. The embryo is ~50 µm long, and the spindle poles and spindle microtubules are clearly visible, whereas astral microtubules are less apparent in these confocal slices. The spindle is initially in the cell centre (top panel) but is displaced towards the posterior during anaphase (middle), resulting in asymmetric cleavage (bottom). (b) Time-lapse fluorescence microscopy of Drosophila sensory organ precursor cell expressing a GFP–tau fusion to mark microtubules and GFP–Pon to mark the axis of polarity. The cell diameter is ~10 µm, and spindle poles are visible, whereas individual microtubules cannot be distinguished. The spindle is initially at a slight angle to the cortical Pon crescent (top panel) but is reoriented so that the spindle is lined up with the polarity axis at cell division (bottom). Photographs from panel (b) reproduced, with permission, from Ref. [50].

Spindle positioning in one-cell-stage C. elegans embryos

Recent experiments demonstrate that unbalanced pulling forces establish proper spindle position along the anteroposterior (AP) axis of one-cell-stage C. elegans embryos. In this organism, a sperm component provides an initial polarizing cue, which probably corresponds to astral microtubules nucleated from the paternally contributed centrosome [25,26]. The sperm component directs six PAR proteins (for ‘partitioning-defective’), as well as the small G protein CDC-42 and the atypical protein kinase C PKC-3, to establish AP polarity [27,28]. In response to polarity cues, the spindle is displaced towards the posterior during anaphase (Fig. 3a). As a result, the first division is unequal, giving rise to a larger anterior blastomere (AB) and a smaller posterior blastomere (P1). In most par mutant embryos, AP polarity cues are not established and posterior spindle displacement does not take place. As a result, the first division is equal.

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

α subunits in Drosophila and C. elegans Box 2. Distinct requirements for Gα In the fruit fly Drosophila melanogaster, the heterotrimeric G protein subunit Gαi was identified biochemically as a binding partner for Pins (Partner of Inscuteable), a GoLoco-domain-containing protein required for spindle orientation and asymmetric distribution of cell-fate determinants in neuroblasts [a,b]. Inactivation of Gαi, which can be achieved using a mutation in the interacting Gβ subunit Gβ13F, results in spindle positioning defects in both neuroblasts and sensory organ precursors [c]. However, these defects are accompanied by dramatic alterations in the distribution of polarity cues. For instance, instead of being localized in a basal cortical crescent as in wild-type neuroblasts, Miranda is delocalized to the cytoplasm or present in a crescent that is not basal. Similarly, instead of being localized in a cortical crescent as in wild-type sensory organ precursors, Numb is distributed around the cell cortex. In the nematode Caenorhabditis elegans, by contrast, the Gα subunits GOA-1 and GPA-16, and the GoLoco-domain-containing proteins AGS-3.2 and AGS-3.3 appear to act downstream of, or in parallel to, polarity cues (main text). Polarity defects might become apparent in C. elegans embryos that lack goa-1/gpa-16 or ags-3.2/ags-3.3 function when additional markers are examined and other cell types investigated. Alternatively, the difference might stem from the different nature of the proteins involved. Indeed, C. elegans GOA-1 and GPA-16 are more closely related to Drosophila Gαo, a Gα subunit that does not bind to the Pins complex [c], than they are to Drosophila Gαi. Moreover, C. elegans AGS-3.2 and AGS-3.3 shows little homology to Drosophila Pins outside the GoLoco domain. Because of these differences, the targets of Gα signaling might be distinct in different organisms and include polarity cues only in Drosophila. References a Schaefer, M. et al. (2000) A protein complex containing Inscuteable and the Gα-binding protein Pins orients asymmetric cell divisions in Drosophila. Curr. Biol. 10, 353–362 b Yu, F. et al. (2000) Analysis of Partner of Inscuteable, a novel player of Drosophila asymmetric divisions, reveals two distinct steps in Inscuteable apical localization. Cell 100, 399–409 c Schaefer, M. et al. (2001) Heterotrimeric G proteins direct two modes of asymmetric cell division in the Drosophila nervous system. Cell 107, 183–194

Spindle severing experiments demonstrate that extra-spindle pulling forces act on spindle poles in wild-type one-cell-stage embryos [29]. Interestingly, the posterior spindle pole moves faster than the anterior one after spindle severing, indicating that a larger net pulling force acts on the posterior spindle pole. Additional spindle severing experiments conducted in par-2 and par-3 mutant embryos establish that the extent of the pulling force is regulated by AP polarity cues. Therefore, polarity translates into asymmetric spindle positioning by modulating the net pulling force acting on each spindle pole. The mechanical basis of these pulling forces remains to be elucidated. It will be interesting to determine whether they are coupled to microtubule depolymerization or driven by cytoplasmic dyneinmediated minus-end-directed motility. Equally interesting will be to discover how polarity cues modulate force-generation mechanisms to result in unbalanced pulling forces. Mathematical modeling suggests that increased pulling forces on the posterior spindle pole might result from the destabilization of interactions between astral microtubules and cortical anchors specifically at the posterior cortex [29]. Two α subunits of heterotrimeric G proteins are important for generating the pulling forces that act on both anterior and posterior spindle poles. The Gα subunits GOA-1 and GPA-16 are required together to promote posterior spindle displacement in one-cell-stage C. elegans embryos [30]. There is a http://tcb.trends.com

335

similar requirement for two essentially identical proteins tentatively named AGS-3.2 and AGS-3.3 (after the distantly related mammalian activator of G-protein signaling 3), which harbour a GoLoco domain characteristic of molecules that interact with Gα subunits [31,32]. The distributions of PAR proteins and of P granules, which are normally segregated to the posterior of one-cell-stage embryos in response to AP polarity [33,34], are not affected in goa-1/gpa-16 (RNAi) or ags-3.2/ags-3.3 (RNAi) embryos, indicating that polarity cues are intact [30] (K. Colombo and P. Gönczy, unpublished). Strikingly, spindle severing experiments reveal that pulling forces acting on both spindle poles are dramatically decreased in such embryos (P. Gönczy and S. Grill, unpublished). Therefore, GOA-1/GPA-16 and AGS-3.2/AGS-3.3 are essential for force generation in one-cell-stage C. elegans embryos. How this is achieved remains to be elucidated. Mammalian Gα subunits can promote microtubule catastrophes in vitro [35], raising the possibility that Gα subunits in one-cell-stage C. elegans embryos promote depolymerization-coupled movements and the generation of pulling forces. Interestingly, although a Gα subunit and a GoLocodomain-containing protein are also required for proper spindle positioning in Drosophila, they act at the level of polarity cues in that organism (Box 2) [36–38]. Thus, signaling via α subunits of heterotrimeric G proteins appears to play distinct roles in governing asymmetric cell division in C. elegans and in Drosophila. Rotation events that orient the spindle in C. elegans and in Drosophila

Interactions between a spatially restricted cortical site and astral microtubules ensure proper spindle orientation in the P1 blastomere of two-cell-stage C. elegans embryos. In this cell, the centrosome pair lies initially perpendicular to the AP axis. During prophase, however, one centrosome moves towards a site in the anterior cortex, thus inducing a 90° rotation of the associated nucleus and other centrosome. As a result, the spindle sets up along the AP axis, and factors such as P granules that are segregated to the posterior of P1 are inherited strictly by its posterior daughter [39]. Local laser irradiation experiments indicate that the 90° rotation is driven by interactions between astral microtubules emanating from the leading centrosome and a spatially restricted cortical site [40]. What is the nature of this cortical site? Components of the dynactin complex and the dynein heavy chain are enriched at that location [41–43], which suggests a model in which cortically anchored dynein interacts with astral microtubules and pulls the associated centrosome by minus-enddirected motility, thus inducing rotation. Although it is attractive, this model has not been directly tested, because embryos lacking dynein or dynactin function after inactivation by RNA-mediated interference have defects in the one-cell stage that preclude assessment of a potential requirement in P1 [41,42].

336

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

Cell cortex

β-cat

APC EB-1

α-cat +

–

TRENDS in Cell Biology

Fig. 4. Hypothetical model of interaction between cortical anchors and astral microtubules. α-Catenin (α-cat), β-catenin (β-cat) and the adenomatous polyposis coli protein (APC) form a complex that binds to cortical F-actin through the α-catenin moiety. The complex is linked to astral microtubules through the interaction between APC and EB-1, which is located at the plus ends of microtubules.

One crucial component of cortical anchor function in C. elegans might be let-99 (for ‘lethal’) [44]. let-99 mutant embryos exhibit spindle positioning defects in many embryonic blastomeres. For instance, the characteristic 90° rotation does not take place in P1. Importantly, polarity cues are not affected in let-99 mutant embryos, because asymmetric distribution of PAR-3, PAR-2, PAR-1 and P granules is the same as in the wild type. Therefore, let-99 is required downstream of, or in parallel to, polarity cues to mediate proper spindle positioning. How does let-99 achieve its function? Time-lapse microscopy shows that nucleus–centrosome complexes in let-99 mutant embryos undergo abnormal vigorous movements towards and away from the cell cortex, suggesting that interactions between astral microtubules and cortical anchors are established but not properly maintained. The let-99 gene encodes a protein with no clear ortholog in other organisms, and it will be particularly interesting to elucidate how it might modulate pulling forces that act on centrosomes and spindle poles. Rotations events similar to the one observed in the P1 blastomere of C. elegans embryos are crucial for achieving proper spindle orientation in Drosophila neuroblasts and sensory organ precursor cells. During development of the fly’s central nervous system, neuroblasts divide asymmetrically in a stem-cell lineage, generating a neuroblast and a smaller ganglion mother cell at each cell division [45]. Dedicated proteins destined to the ganglion mother cell, including Numb, Partner of Numb (Pon), Prospero and Miranda, become localized during mitosis in a crescent on the basal side of neuroblasts [46,47]. For these components to be correctly partitioned at cell division, the spindle must be positioned along the apico-basal axis. Live imaging of a fusion protein between green fluorescent protein (GFP) and the microtubule-associated protein tau reveals that the centrosome pair initially lies perpendicular to this axis [48]. At the onset of mitosis, http://tcb.trends.com

however, the spindle poles undergo a 90° rotation, thus achieving proper spindle orientation by anaphase. Rotation also positions the spindle along an axis of polarity in the Drosophila sensory organ precursor cell, which similarly undergoes asymmetric division [49]. In this case, live imaging using GFP–tau and GFP–Pon demonstrates that spindle orientation is achieved so as to align the spindle with the asymmetrically distributed Pon crescent (Fig. 3b) [50]. Cortical anchors in metazoan organisms

Although the molecular nature of cortical anchors underlying rotation events in C. elegans and Drosophila remains to be elucidated, recent findings suggest that an evolutionarily conserved complex containing a protein related to adenomatous polyposis coli (APC) can link cortical actin to astral microtubules (Fig. 4). In the syncytial Drosophila embryo, the β-catenin homolog Armadillo (Arm) is required to tether the spindle to cortical actin [51]. Mutant analysis indicates that a complex composed of Arm, α-catenin and the APC-related protein dAPC2 interacts with cortical actin through the α-catenin moiety. Interference with dAPC2 function later in development results in spindle positioning defects in neurectodermal cells, and similar defects are observed following inactivation of dEB1, a Drosophila homolog of EB1 [52]. Interestingly, human EB1 is a microtubuleassociated protein that localizes to the plus ends of growing microtubules and interacts with APC [53,54]. Although dEB1 does not appear to interact with dAPC2 [52], the presence of two EB1-related genes and two APC-related genes in Drosophila leaves open the possibility that the molecular interaction between an α-catenin–β-catenin–APC complex and EB1 is conserved across metazoan evolution. In fact, the functional conservation might extend in part to unicellular eukaryotes, because S. cerevisiae Kar9p has a small stretch of homology with APC and interacts with the EB1 homolog Bim1p [19–21,55]. Although a complex containing APC/Kar9p appears to play an important role in some instances of spindle positioning, distinct cortical anchors for astral microtubules probably remain to be discovered. Relationships between polarity cues and spindle positioning

Spindle position must be coordinated with cell polarity to ensure the correct segregation of cell fate determinants during development (Fig. 1b). Such coordination is achieved chiefly because polarity cues control spindle positioning. Thus, the spindle is not displaced towards the posterior of one-cell-stage C. elegans embryos in most par mutants [34]. Likewise, spindle orientation is impaired in Drosophila inscuteable mutant neuroblasts, which lack proper apico-basal polarity [56]. In some cases, polarity cues might control spindle positioning quite directly. For example, par-1 encodes a Ser/Thr protein

Review

(a) Wild-type

TRENDS in Cell Biology Vol.12 No.7 July 2002

(b) spn-4 mutant

TRENDS in Cell Biology

Fig. 5. Relationship between polarity cues and spindle positioning in the P1 blastomere of wild-type and spn-4 mutant Caenorhabditis elegans embryos. Spindle poles are shown as black disks, astral and spindle microtubules as black lines, nuclei as gray disks, and the cleavage furrow as a dashed red line. PAR-2 is shown as a thick green line, P granules as blue disks and the cortical anchor as an orange square. (a) In wild-type embryos, PAR-2 becomes restricted to the posterior cortex and P granules are segregated to the posterior of the P1 blastomere. The characteristic 90° rotation of the centrosome–nucleus complex takes place after this. Therefore, the spindle is established along the longitudinal axis. As a result, PAR-2 and P granules will be present only in the posterior daughter cell P2. (b) In spn-4 mutant embryos, as in wild-type, PAR-2 becomes restricted to the posterior cortex, whereas P granules are segregated to the posterior of the P1 blastomere. However, the 90° rotation of the centrosome–nucleus complex does not take place and so the spindle is established perpendicular to the longitudinal axis. PAR-2 and P granules relocalize before mitosis so as to lie above one of the spindle poles, thus restoring their spatial relationship to the spindle. As a result, PAR-2 and P granules will be present in only one daughter cell, as in the wild-type.

kinase whose mammalian ortholog can phosphorylate microtubule-associated proteins, leading to microtubule destabilization [57,58]. Because PAR-1 is restricted to the posterior cortex of one-cell-stage C. elegans embryos, it is tempting to speculate that its activity leads to local destabilization of microtubules, thus generating depolymerization-coupled movements and increased pulling forces. Compatible with a role in modulating microtubule dynamics, reduction of par-1 function in Drosophila oocytes results in a disorganized microtubule network, as well as in defective AP axis formation [59,60]. In other cases, polarity cues might control spindle positioning less directly, perhaps by modulating components of the force generation machinery or cortical anchors. Coordination between spindle position and cell polarity might, in some cases, be also achieved by http://tcb.trends.com

337

altering the localization of polarity cues or cell-fate determinants to align them with a mispositioned spindle. Such feedback alterations might take place in spn-4 (for ‘spindle orientation defective’) mutant embryos of C. elegans (Fig. 5) [61]. The spn-4 gene encodes a predicted RNA-binding protein of the RRM family and is required for the characteristic 90° rotation in P1. In spn-4 mutant embryos, PAR-2 and P granules are initially restricted to the posterior of P1, as in the wild type, indicating that the lack of rotation does not result from general polarity defects. Surprisingly, however, PAR-2 and P granules relocalize above one of the presumptive spindle poles before mitosis. In this manner, PAR-2 and P granules are segregated to only one daughter cell, as in the wild type, despite the absence of rotation. Although it is possible that the lack of spn-4 function affects polarity cues strictly before mitosis, an attractive alternative is that PAR-2 and P granules relocalize in response to the spindle-positioning defect. A potentially similar situation is encountered in Drosophila neuroblasts with the asterless mutation [62]. In these mutants, Miranda cortical crescents are incorrectly localized in 52% of cells in metaphase, but only in 11% of cells in telophase. Although polarity cues and cell-fate determinants are correctly localized in the complete absence of spindle in C. elegans embryos and Drosophila neuroblasts [33,41,63], the above observations suggest that the distribution of polarity cues and cell fate determinants might be altered in some cases to align them with a mispositioned mitotic spindle. Spindle positioning checkpoint in metazoan organisms?

As an extension to the observations discussed above, it is worth considering whether there might be a spindle positioning checkpoint in metazoan organisms. The existence of such a checkpoint is well established in S. cerevisiae, in which genetic and biochemical evidence supports a model in which Bub2p, Tem1p and Lte1p play a central role in monitoring spindle position [16,64]. Both Bub2p and Tem1p are associated preferentially with the spindle pole body that is destined for the bud, and Bub2p is a component of a GTPase-activating protein (GAP) that inactivates the Tem1p GTPase. By contrast, Lte1p is present exclusively in the bud and is a GTPaseexchange factor (GEF) that activates Tem1p. It is thought that mitotic exit is prevented by Bub2p until Tem1p comes into contact with Lte1p when the spindle is correctly positioned in the bud neck. The available evidence does not support the existence of a strict spindle positioning checkpoint in C. elegans or Drosophila. Thus, cell division proceeds with apparently normal kinetics in embryos lacking goa-1/gpa-16 or let-99 function, despite severe spindle positioning defects [30,44]. However, the spindle assembly checkpoint is not functional either in early embryos of most organisms, raising the

338

Acknowledgments I am grateful to B. Bowerman, J. Knoblich, M. Peter and T. Surrey for critical reading of the manuscript, and to Y Bellaïche for providing images for Fig. 4b. Work in my laboratory is supported by the Swiss National Foundation, the Swiss Cancer League and the EMBO Young Investigator Programme.

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

possibility that a spindle positioning checkpoint might yet be discovered in somatic cells of C. elegans or Drosophila. Micromanipulation experiments in mammalian cells show that a misoriented spindle can delay anaphase onset [65]. It will be interesting to investigate to what extent a spindle positioning checkpoint may be widespread in metazoan organisms to ensure coordination of spatial and temporal aspects of cell division. Concluding remarks

I have discussed a mechanism for spindle positioning in which interactions between cortical anchors and the plus ends of microtubules result in the generation of pulling forces acting on spindle poles. Exciting questions lie ahead. It will be important to elucidate the mechanical basis of force

References 1 Rappaport, R. (1971) Cytokinesis in animal cells. Int. Rev. Cytol. 31, 169–213 2 Holy, T.E. et al. (1997) Assembly and positioning of microtubule asters in microfabricated chambers. Proc. Natl. Acad. Sci. U. S. A. 94, 6228–6231 3 Shaw, S.L. et al. (1997) Astral microtubule dynamics in yeast: a microtubule-based searching mechanism for spindle orientation and nuclear migration into the bud. J. Cell Biol. 139, 985–994 4 Lutz, D.A. et al. (1988) Micromanipulation studies of the asymmetric positioning of the maturation spindle in Chaetopterus sp. oocytes. I. Anchorage of the spindle to the cortex and migration of a displaced spindle. Cell. Motil. Cytoskeleton 11, 83–96 5 Aist, J.R. et al. (1993) Astral and spindle forces in PtK2 cells during anaphase B: a laser microbeam study. J. Cell Sci. 104, 1207–1216 6 Pellman, D. (2001) Cancer. A CINtillating new job for the APC tumor suppressor. Science 291, 2555–2556 7 Bloom, K. (2000) It’s a kar9ochore to capture microtubules. Nat. Cell Biol. 2, E96–E98 8 Mitchison, T.J. and Salmon, E.D. (2001) Mitosis: a history of division. Nat. Cell Biol. 3, E17–E21 9 Lombillo, V.A. et al. (1995) Minus-end-directed motion of kinesin-coated microspheres driven by microtubule depolymerization. Nature 373, 161–164 10 Pfarr, C.M. et al. (1990) Cytoplasmic dynein is localized to kinetochores during mitosis. Nature 345, 263–265 11 Steuer, E.R. et al. (1990) Localization of cytoplasmic dynein to mitotic spindles and kinetochores. Nature 345, 266–268 12 Sharp, D.J. et al. (2000) Cytoplasmic dynein is required for poleward chromosome movement during mitosis in Drosophila embryos. Nat. Cell Biol. 2, 922–930 13 Busson, S. et al. (1998) Dynein and dynactin are localized to astral microtubules and at cortical sites in mitotic epithelial cells. Curr. Biol. 8, 541–544 14 Han, G. et al. (2001) The Aspergillus cytoplasmic dynein heavy chain and NUDF localize to microtubule ends and affect microtubule dynamics. Curr. Biol. 11, 719–724 15 Carminati, J.L. and Stearns, T. (1997) Microtubules orient the mitotic spindle in yeast through dynein-dependent interactions with the cell cortex. J. Cell Biol. 138, 629–641

http://tcb.trends.com

generation and to determine the molecular nature of cortical anchors. It will also be interesting to discover how cortical anchors and minus-end-directed motors such as cytoplasmic dynein can maintain interactions with depolymerizing microtubules, so as to generate pulling forces on spindle poles. Another major challenge lies in understanding how temporal and spatial regulation of microtubule dynamics might be used, perhaps under the control of polarity cues, to modulate the extent of pulling forces acting on spindle poles. Spindle position in some cells is influenced by cell-extrinsic factors and it will be of interest to unravel how the core machinery is altered in these cases. These and other advances will generate a more comprehensive understanding of the mechanisms underlying spatial cell division control.

16 Schuyler, S.C. and Pellman, D. (2001) Search, capture and signal: games microtubules and centrosomes play. J. Cell Sci. 114, 247–255 17 Segal, M. and Bloom, K. (2001) Control of spindle polarity and orientation in Saccharomyces cerevisiae. Trends Cell Biol. 11, 160–166 18 Miller, R.K. and Rose, M.D. (1998) Kar9p is a novel cortical protein required for cytoplasmic microtubule orientation in yeast. J. Cell Biol. 140, 377–390 19 Lee, L. et al. (2000) Positioning of the mitotic spindle by a cortical-microtubule capture mechanism. Science 287, 2260–2262 20 Miller, R.K. et al. (2000) Bim1p/Yeb1p mediates the Kar9p-dependent cortical attachment of cytoplasmic microtubules. Mol. Biol. Cell 11, 2949–2959 21 Korinek, W.S. et al. (2000) Molecular linkage underlying microtubule orientation toward cortical sites in yeast. Science 287, 2257–2259 22 Farkasovsky, M. and Kuntzel, H. (2001) Cortical Num1p interacts with the dynein intermediate chain Pac11p and cytoplasmic microtubules in budding yeast. J Cell Biol 152, 251–262 23 Farkasovsky, M. and Kuntzel, H. (1995) Yeast Num1p associates with the mother cell cortex during S/G2 phase and affects microtubular functions. J. Cell Biol. 131, 1003–1014 24 Heil-Chapdelaine, R.A. et al. (2000) The cortical protein Num1p is essential for dynein-dependent interactions of microtubules with the cortex. J. Cell Biol. 151, 1337–1344 25 Wallenfang, M.R. and Seydoux, G. (2000) Polarization of the anterior–posterior axis of C. elegans is a microtubule-directed process. Nature 408, 89–92 26 Goldstein, B. and Hird, S.N. (1996) Specification of the anteroposterior axis in Caenorhabditis elegans. Development 122, 1467–1474 27 Kemphues, K.J. and Strome, S. (1997) Fertilization and establishment of polarity in the embryo. In C. elegans (Vol. II) (Riddle, D.L. et al., eds), pp. 335–359, Cold Spring Harbor Laboratory Press 28 Gotta, M. and Ahringer, J. (2001) Axis determination in C. elegans: initiating and transducing polarity. Curr. Opin. Genet. Dev. 11, 367–373 29 Grill, S.W. et al. (2001) Polarity controls forces governing asymmetric spindle positioning in the Caenorhabditis elegans embryo. Nature 409, 630–633

30 Gotta, M. and Ahringer, J. (2001) Distinct roles for Gα and Gβγ in regulating spindle position and orientation in Caenorhabditis elegans embryos. Nat. Cell Biol. 3, 297–300 31 Siderovski, D.P. et al. (1999) The GoLoco motif: a Gαi/o binding motif and potential guaninenucleotide exchange factor. Trends Biochem. Sci. 24, 340–341 32 Gönczy, P. et al. (2000) Functional genomic analysis of cell division in C. elegans using RNAi of genes on chromosome III. Nature 408, 331–336 33 Strome, S. and Wood, W.B. (1983) Generation of asymmetry and segregation of germ-line granules in early C. elegans embryos. Cell 35, 15–25 34 Kemphues, K.J. et al. (1988) Maternal-effect lethal mutations on linkage group II of Caenorhabditis elegans. Genetics 120, 977–986 35 Roychowdhury, S. et al. (1999) G protein alpha subunits activate tubulin GTPase and modulate microtubule polymerization dynamics. J. Biol. Chem. 274, 13485–13490 36 Schaefer, M. et al. (2000) A protein complex containing Inscuteable and the Gα-binding protein Pins orients asymmetric cell divisions in Drosophila. Curr. Biol. 10, 353–362 37 Yu, F. et al. (2000) Analysis of Partner of Inscuteable, a novel player of Drosophila asymmetric divisions, reveals two distinct steps in Inscuteable apical localization. Cell 100, 399–409 38 Schaefer, M. et al. (2001) Heterotrimeric G proteins direct two modes of asymmetric cell division in the Drosophila nervous system. Cell 107, 183–194 39 Hyman, A.A. and White, J.G. (1987) Determination of cell division axes in the early embryogenesis of Caenorhabditis elegans. J. Cell Biol. 105, 2123–2135 40 Hyman, A.A. (1989) Centrosome movement in the early divisions of Caenorhabditis elegans: a cortical site determining centrosome position. J. Cell Biol. 109, 1185–1193 41 Gönczy, P. et al. (1999) Cytoplasmic dynein is required for distinct aspects of MTOC positioning, including centrosome separation, in the one cell stage Caenorhabditis elegans embryo. J. Cell Biol. 147, 135–150 42 Skop, A.R. and White, J.G. (1998) The dynactin complex is required for cleavage plane specification in early Caenorhabditis elegans embryos. Curr. Biol. 8, 1110–1116

Review

TRENDS in Cell Biology Vol.12 No.7 July 2002

43 Waddle, J.A. et al. (1994) Transient localized accumulation of actin in Caenorhabditis elegans blastomeres with oriented asymmetric divisions. Development 120, 2317–2328 44 Rose, L.S. and Kemphues, K. (1998) The let-99 gene is required for proper spindle orientation during cleavage of the C. elegans embryo. Development 125, 1337–1346 45 Doe, C.Q. and Skeath, J.B. (1996) Neurogenesis in the insect central nervous system. Curr. Opin. Neurobiol. 6, 18–24 46 Doe, C.Q. and Bowerman, B. (2001) Asymmetric cell division: fly neuroblast meets worm zygote. Curr. Opin. Cell Biol. 13, 68–75 47 Schaefer, M. and Knoblich, J.A. (2001) Protein localization during asymmetric cell division. Exp. Cell Res. 271, 66–74 48 Kaltschmidt, J.A. et al. (2000) Rotation and asymmetry of the mitotic spindle direct asymmetric cell division in the developing central nervous system. Nat. Cell Biol. 2, 7–12 49 Bellaiche, Y. and Schweisguth, F. (2001) Lineage diversity in the Drosophila nervous system. Curr. Opin. Genet. Dev. 11, 418–423 50 Bellaiche, Y. et al. (2001) Frizzled regulates localization of cell-fate determinants and mitotic spindle rotation during asymmetric cell division. Nat. Cell Biol. 3, 50–57

51 McCartney, B.M. et al. (2001) Drosophila APC2 and Armadillo participate in tethering mitotic spindles to cortical actin. Nat. Cell Biol. 3, 933–938 52 Lu, B. et al. (2001) Adherens junctions inhibit asymmetric division in the Drosophila epithelium. Nature 409, 522–525 53 Su, L.K. et al. (1995) APC binds to the novel protein EB1. Cancer Res. 55, 2972–2977 54 Mimori-Kiyosue, Y. et al. (2000) The dynamic behavior of the APC-binding protein EB1 on the distal ends of microtubules. Curr. Biol. 10, 865–868 55 Bienz, M. (2001) Spindles cotton on to junctions, APC and EB1. Nat Cell Biol 3, E67–E68 56 Kraut, R. et al. (1996) Role of Inscuteable in orienting asymmetric cell divisions in Drosophila. Nature 383, 50–55 57 Guo, S. and Kemphues, K.J. (1995) par-1, a gene required for establishing polarity in C. elegans embryos, encodes a putative Ser/Thr kinase that is asymmetrically distributed. Cell 81, 611–620 58 Drewes, G. et al. (1997) MARK, a novel family of protein kinases that phosphorylate microtubule-associated proteins and trigger microtubule disruption. Cell 89, 297–308

339

59 Shulman, J.M. et al. (2000) The Drosophila homolog of C. elegans PAR-1 organizes the oocyte cytoskeleton and directs oskar mRNA localization to the posterior pole. Cell 101, 377–388 60 Tomancak, P. et al. (2000) A Drosophila melanogaster homologue of Caenorhabditis elegans par-1 acts at an early step in embryonic-axis formation. Nat. Cell Biol. 2, 458–460 61 Gomes, J.E. et al. (2001) The maternal gene spn-4 encodes a predicted RRM protein required for mitotic spindle orientation and cell fate patterning in early C. elegans embryos. Development 128, 4301–4314 62 Giansanti, M.G. et al. (2001) The role of centrosomes and astral microtubules during asymmetric division of Drosophila neuroblasts. Development 128, 1137–1145 63 Knoblich, J.A. et al. (1995) Asymmetric segregation of Numb and Prospero during cell division. Nature 377, 624–627 64 Pereira, G. and Schiebel, E. (2001) The role of the yeast spindle pole body and the mammalian centrosome in regulating late mitotic events. Curr. Opin. Cell Biol. 13, 762–769 65 O’Connell, C.B. and Wang, Y.L. (2000) Mammalian spindle orientation and position respond to changes in cell shape in a dynein-dependent fashion. Mol. Biol. Cell 11, 1765–1774

Ephrins in reverse, park and drive Chad A. Cowan and Mark Henkemeyer Eph receptors and their membrane-anchored ephrin ligands are thought to orchestrate cell movements by transducing bidirectional tyrosine-kinasemediated signals into both cells expressing the receptors and cells expressing the ligands. Whether the resulting event is repulsion of an axonal growth cone, directing the orderly segmentation of hindbrain rhombomere cells or controlling angiogenic remodelling, such elaborate and diverse cell movements require intricate changes in the actin cytoskeleton, as well as precise regulation of cellular adhesion. Recent work by several groups has begun to link ephrin reverse signals to intracellular pathways that regulate actin dynamics and might help to explain how these ligands function as receptors to direct cell movement, adhesion and de-adhesion events.

Chad A. Cowan* Mark Henkemeyer† Center for Developmental Biology and Kent Waldrep Foundation Center for Basic Research on Nerve Growth and Regeneration, University of Texas Southwestern Medical Center, Dallas, TX 75390-9133, USA. e-mails: *cowan@ utsw.swmed.edu †[email protected]

The development of multicellular organisms requires an exquisite interplay of cell proliferation, differentiation, adherence and movement. Fundamental to normal development are many different cell-anchored receptors that communicate information into their cell after exposure to the appropriate ligand. Whether soluble and acting at a distance, or anchored to the membrane and acting on cell–cell contact, most ligand–receptor systems transduce information unidirectionally from the ligand-expressing cell into the receptorexpressing cell. Presenting a contrast to this model are the large family of Eph receptor tyrosine kinases and their http://tcb.trends.com

membrane-anchored ephrin ligands. On cell–cell contact, these molecules transduce important signals bidirectionally into both the receptor-expressing cell and the ligand-expressing cell in what is known as ‘forward’ and ‘reverse’ signaling, respectively [1] (Fig. 1). In this way, the Eph receptors can also function as ligands and the ephrin ligands can also function as receptors. Bidirectional signaling mediated by Ephs and ephrins is thought to communicate signals that regulate the cytoskeleton during axon pathfinding, cell migration, cell adhesion and vascular remodelling (reviewed in [2–5]) and is now also linked to the regulation of fluid production and ion homeostasis in nonmotile epithelial cells [6]. As the known range of biological functions performed by this family of multitalented receptors and ligands continues to expand, focus is shifting to determining how these molecules transduce such important and diverse signals into the cell. Ephs and ephrins can be divided into two classes on the basis of their structural features and binding affinities. A-subclass ephrins (ephrin-A1 to ephrin-A5) are attached to the outer leaflet of the plasma membrane by a glycosylphosphatidylinositol (GPI) linkage. They bind to and activate the A-subclass Eph receptors

0962-8924/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S0962-8924(02)02317-6